Breakout datasets: HUD-masked collection, full-HUD Checkpoint Monitoring Runs, and derived world models.
🔄 In a Training Loop
Tiago Silva
tsilva
AI & ML interests
Deep Reinforcement Learning
Recent Activity
updated a dataset about 16 hours ago
tsilva/banking77-label-quality updated a dataset about 19 hours ago
tsilva/scummhd-queen published a dataset about 19 hours ago
tsilva/scummhd-queenOrganizations
Historical rlab / Stable-Baselines3 PPO Policies, plus a separate gymrec recording. These are not GradLab-trained Checkpoints.
-
tsilva/Level1-1_stable-baselines3-ppo_7d9b9131
Reinforcement Learning • Updated • 33 -
tsilva/Level1-2_stable-baselines3-ppo_91df110f
Reinforcement Learning • Updated • 15 -
tsilva/Level1-4_stable-baselines3-ppo_be9abbc8
Reinforcement Learning • Updated • 8 -
tsilva/Level2-1_stable-baselines3-ppo_4792f358
Reinforcement Learning • Updated • 8
Breakout datasets: HUD-masked collection, full-HUD Checkpoint Monitoring Runs, and derived world models.
Evidence-backed GradLab research releases for VizdoomDeathmatch-v1.
Historical rlab / Stable-Baselines3 PPO Policies, plus a separate gymrec recording. These are not GradLab-trained Checkpoints.
-
tsilva/Level1-1_stable-baselines3-ppo_7d9b9131
Reinforcement Learning • Updated • 33 -
tsilva/Level1-2_stable-baselines3-ppo_91df110f
Reinforcement Learning • Updated • 15 -
tsilva/Level1-4_stable-baselines3-ppo_be9abbc8
Reinforcement Learning • Updated • 8 -
tsilva/Level2-1_stable-baselines3-ppo_4792f358
Reinforcement Learning • Updated • 8
MNIST classification using different recipes.